Spotting AI fake news before you forward it

AI makes false posts, pictures and voice notes easy to produce. A simple stop-and-check routine helps you avoid spreading them.

Why this matters now

False stories have always spread on social media. AI has made them cheaper and faster to produce. One person can now generate many convincing posts, fake screenshots, cloned voice notes and edited videos in a single afternoon.

Some of this is misinformation: false content shared by people who believe it. Some is disinformation: false content made on purpose to mislead, cheat or stir up anger. Either way, the person who forwards it helps it spread.

You cannot check everything. You can build a habit of pausing before you share.

Where it shows up

WhatsApp forwards marked as forwarded many times, often with no source, sent into family groups
Facebook posts and pages with dramatic headlines, made-up quotes or pictures of public figures
TikTok and other short videos where a real clip is given a false caption. Sometimes a voice is added that the person never recorded
Posts on X that look like breaking news from an official account but come from a copycat profile

Common tricks

Fake screenshots of news channels, official notices or social media posts. These are easy to make and need no AI at all.
Made-up quotes placed on a photo of a politician, scholar, cricketer or celebrity.
AI-written "news" articles on websites with names that sound like real newspapers.
Old photos or videos reused with a new caption about a current event.
Cloned voice notes claiming to be from a doctor, official or religious figure. See deepfake and voice scams.
Health and money rumours: miracle cures, scary warnings about medicines, sudden changes to bank rules, free money schemes.

The stop-and-check routine

Before you forward anything that surprises, frightens or angers you, take one minute.

Stop. Strong emotion is the signal. False posts are designed to make you react quickly.
Ask who made it. Is there a named source, such as a real news organisation, office or person? "Received as forwarded" is not a source.
Look for the same story elsewhere. If something big has happened, more than one established news outlet will report it. Search for the main claim in your own words.
Check official channels directly. For a claim about a government decision, look at that office's own website or verified social media account. Do not trust a screenshot of it.
Look closely at pictures and video. Check whether the image has appeared before with a different story. See AI images and videos.
When in doubt, do not forward. Not sharing costs you nothing.

Warning signs

No date, no place and no named source
Phrases like "share before it is deleted" or "the media will not show you this"
A misspelt name of an official body
A logo or letterhead that looks slightly wrong
A link to a website you have never heard of
A request for money, personal details or a click on a link
Content that confirms exactly what you already fear or believe

Can a chatbot check it for me?

A chatbot can help you think through a message and point out warning signs. Try our scam check prompt. But chatbots can be wrong, out of date or simply make things up. This is called hallucination. Use a chatbot as a second opinion, never as proof that something is true.

If you have already shared it

Delete the message from the group if you can.
Send a short follow-up: "I shared this earlier. It appears to be false. Please do not forward it."
Use the app's report option for posts that are clearly fake or harmful.

Correcting yourself is a sign of good judgement. Others in the group will learn from it too.

For what happens to what you type into AI tools, see AI and your privacy.